Dataversal Sensitivity

Raspberry Pi system can detect viruses on other devices without use of software

Jan 2022, phys.org

https://techxplore.com/news/2022-01-raspberry-pi-viruses-devices-software.html

A Raspberry Pi, an H-field probe and an oscilloscope.

Software generates electromagnetic waves -- which basically means that every piece of software has a unique fingerprint of wave patterns that can be used to detect it. But guess what, so do viruses. So as long as we know what the virus is beforehand, we can put its fingerprints (waveprints?) on file, and match against it. 

No downloading required -- this new approach only needs to be close enough to the computer to detect the waves, from outside the machine.

It's not directly related but this does remind me of the neural networked doppler vibrometer that can inventory all the mechanical equipment in your house by the minute, subtle vibration patterns they make on your ceiling. 

via Institute of Computer Science and Random Systems in France: Duy-Phuc Pham et al, Obfuscation Revealed: Leveraging Electromagnetic Signals for Obfuscated Malware Classification, Annual Computer Security Applications Conference (2021). DOI: 10.1145/3485832.3485894

http://dx.doi.org/10.1145/3485832.3485894

Image credit: 3D Fractal Caleidscope by Fractal Fertilizer for Fractal Forums

Using artificial intelligence to find anomalies hiding in massive datasets

Feb 2022, phys.org

https://techxplore.com/news/2022-02-artificial-intelligence-anomalies-massive-datasets.html

It's network science -- 

It learns to model the interconnectedness of the power grid using Bayesian-networked, normalizing flow, deep-learning model.

Rule-based systems, even empowered by statistical data analysis, require a lot of labor and expertise, but this is easier to apply in real-world situations where high-quality labeled datasets are often hard to come by. Their method is especially powerful because this complex graph structure does not need to be defined in advance — the model can learn the graph on its own, in an unsupervised manner [no labels required]. Their methodology is also flexible. Armed with a large, unlabeled dataset, they can tune the model to make effective anomaly predictions in other situations, like traffic patterns. They also want to explore how they can efficiently learn these models when the graphs become enormous, perhaps with millions or billions of interconnected nodes.

via MIT-IBM Watson AI Lab: Graph-augmented Normalizing Flows for Anomaly Detection of Multiple Time Series, Enyan Dai, Jie Chen. openreview.net/forum?id=45L_dgP48Vd

https://openreview.net/forum?id=45L_dgP48Vd